Beauty is in the ELBO of the Beholder: A Variational Account of Processing Fluency in Face Perception
Abstract
Facial attractiveness has been linked to statistical regularities such as symmetry and averageness, suggesting that beauty may depend on the ease with which a face is perceived. We empirically test this hypothesis by training variational autoencoders on four face datasets without attractiveness supervision and evaluating their representations on the 597 faces from the Chicago Face Database. We report three main novel findings. First, we identify that human attractiveness ratings closely aligns with the direction defined by the VAE evidence lower bound (ELBO) in rate-distortion space. Second, we find that independently learned latent spaces contain an attractiveness direction that transfers strongly across random initializations and training data. Finally, we observe that attractive faces are more prototypical in both shape and latent space. Our work provides the first explicit connection between classic accounts of aesthetics with learned generative models and empirical support for a variational interpretation of the processing fluency theory of aesthetic pleasure.
Related
- Investigating lexical knowledge and lexical processing in order to explore their role in second language utterance fluency
- Fluently Grooving: Investigating neural and behavioural indices of processing fluency in complex auditory rhythms
- Propositional and imagistic incongruity: a dual-route cognitive account of humor processing
- MEM as an Extension of Recurrent Processing Theory: Toward a Motivational Account of Consciousness
- Frequency-specific oscillations track parallel processing of linguistic structural and lexical-semantic information from perception to decision
BibTeX
@misc{lopez2026beauty,
title = {{Beauty is in the ELBO of the Beholder: A Variational Account of Processing Fluency in Face Perception}},
author = {Francisco M. López and Jochen Triesch},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2608.24219},
eprint = {2608.24219},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.24219},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Beauty is in the ELBO of the Beholder: A Variational Account of Processing Fluency in Face Perception
AU - Francisco M. López
AU - Jochen Triesch
PY - 2026
DA - 2026/10/20
PB - arXiv
DO - 10.48550/arxiv.2608.24219
UR - https://arxiv.org/abs/2608.24219
AB - Facial attractiveness has been linked to statistical regularities such as symmetry and averageness, suggesting that beauty may depend on the ease with which a face is perceived. We empirically test this hypothesis by training variational autoencoders on four face datasets without attractiveness supervision and evaluating their representations on the 597 faces from the Chicago Face Database. We report three main novel findings. First, we identify that human attractiveness ratings closely aligns with the direction defined by the VAE evidence lower bound (ELBO) in rate-distortion space. Second, we find that independently learned latent spaces contain an attractiveness direction that transfers strongly across random initializations and training data. Finally, we observe that attractive faces are more prototypical in both shape and latent space. Our work provides the first explicit connection between classic accounts of aesthetics with learned generative models and empirical support for a variational interpretation of the processing fluency theory of aesthetic pleasure.
ER -